1. Field of the Invention
The present invention relates to an image signal processing apparatus and, more particularly, to an image signal processing apparatus comprising means for digitaizing an image signal.
An electrophotographic copying apparatus and a facsimile terminal equipment read a document and copy its image into a desired form or transmit it. They comprise an image signal processing apparatus for processing a read image signal for those operation. The present invention aims to improve such an image signal processing apparatus.
2. Description of the Conventional Art
A contour extracting process is one of signal processes performed in the image signal processing apparatus. A method for extracting contours of image information is disclosed in Japanese Patent Laying Open Gazette No. 60-59484, as a prior art of interest to the present invention. According to the method described in this gazette as shown in FIG. 12, binarization is performed two times in accordance with two threshold values L1 and L2 when the image information is divided into bilevels and then only a signal component L in an intermediate region is taken out to extract a contour line.
In the above-described image information contour extracting method, it is necessary to set two high and low threshold values L1 and L2 at the time of binarization, but it is difficult to set the threshold values. For example, when the image signal to be binarized is a signal of a character which is very light in color and written on a white background or a signal of a character which is very dark in color and written on a black background, a level difference between the background and the character signals is small. Therefore, if a level difference between high and low threshold values is too large, the image signal can not be binarized at two levels as shown in FIG. 13, so that the contour line can not be extracted. In this case, the level difference between two threshold values may be reduced. However, since the image signal sometimes fluctuates due to a noise and the like, if this level difference is made too small, it is difficult to precisely extract a component in the intermediate region between two threshold values.
In addition, in an image comprising an intermediate tone such as a photograph, the image of the intermediate tone itself is destroyed or erroneously operated when binarized, as shown in FIG. 14.